version 1.118, 2010/03/31 07:42:04 |
version 1.131, 2010/05/14 17:31:25 |
|
|
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
|
#include <time.h> |
|
|
#include "libmdoc.h" |
#include "libmdoc.h" |
#include "libmandoc.h" |
#include "libmandoc.h" |
Line 67 const char *const __mdoc_merrnames[MERRMAX] = { |
|
Line 68 const char *const __mdoc_merrnames[MERRMAX] = { |
|
"line arguments discouraged", /* ENOLINE */ |
"line arguments discouraged", /* ENOLINE */ |
"prologue macro out of conventional order", /* EPROLOOO */ |
"prologue macro out of conventional order", /* EPROLOOO */ |
"prologue macro repeated", /* EPROLREP */ |
"prologue macro repeated", /* EPROLREP */ |
"invalid manual section", /* EBADMSEC */ |
|
"invalid section", /* EBADSEC */ |
"invalid section", /* EBADSEC */ |
"invalid font mode", /* EFONT */ |
"invalid font mode", /* EFONT */ |
"invalid date syntax", /* EBADDATE */ |
"invalid date syntax", /* EBADDATE */ |
Line 83 const char *const __mdoc_merrnames[MERRMAX] = { |
|
Line 83 const char *const __mdoc_merrnames[MERRMAX] = { |
|
"no description found for library", /* ELIB */ |
"no description found for library", /* ELIB */ |
"bad child for parent context", /* EBADCHILD */ |
"bad child for parent context", /* EBADCHILD */ |
"list arguments preceding type", /* ENOTYPE */ |
"list arguments preceding type", /* ENOTYPE */ |
|
"deprecated comment style", /* EBADCOMMENT */ |
}; |
}; |
|
|
const char *const __mdoc_macronames[MDOC_MAX] = { |
const char *const __mdoc_macronames[MDOC_MAX] = { |
Line 140 const char *const __mdoc_argnames[MDOC_ARG_MAX] = { |
|
Line 141 const char *const __mdoc_argnames[MDOC_ARG_MAX] = { |
|
const char * const *mdoc_macronames = __mdoc_macronames; |
const char * const *mdoc_macronames = __mdoc_macronames; |
const char * const *mdoc_argnames = __mdoc_argnames; |
const char * const *mdoc_argnames = __mdoc_argnames; |
|
|
|
static void mdoc_node_free(struct mdoc_node *); |
|
static void mdoc_node_unlink(struct mdoc *, |
|
struct mdoc_node *); |
static void mdoc_free1(struct mdoc *); |
static void mdoc_free1(struct mdoc *); |
static void mdoc_alloc1(struct mdoc *); |
static void mdoc_alloc1(struct mdoc *); |
static struct mdoc_node *node_alloc(struct mdoc *, int, int, |
static struct mdoc_node *node_alloc(struct mdoc *, int, int, |
enum mdoct, enum mdoc_type); |
enum mdoct, enum mdoc_type); |
static int node_append(struct mdoc *, |
static int node_append(struct mdoc *, |
struct mdoc_node *); |
struct mdoc_node *); |
static int parsetext(struct mdoc *, int, char *); |
static int mdoc_ptext(struct mdoc *, int, char *); |
static int parsemacro(struct mdoc *, int, char *); |
static int mdoc_pmacro(struct mdoc *, int, char *); |
static int macrowarn(struct mdoc *, int, const char *); |
static int macrowarn(struct mdoc *, int, const char *); |
static int pstring(struct mdoc *, int, int, |
|
const char *, size_t); |
|
|
|
|
|
const struct mdoc_node * |
const struct mdoc_node * |
mdoc_node(const struct mdoc *m) |
mdoc_node(const struct mdoc *m) |
{ |
{ |
Line 176 mdoc_free1(struct mdoc *mdoc) |
|
Line 179 mdoc_free1(struct mdoc *mdoc) |
|
{ |
{ |
|
|
if (mdoc->first) |
if (mdoc->first) |
mdoc_node_freelist(mdoc->first); |
mdoc_node_delete(mdoc, mdoc->first); |
if (mdoc->meta.title) |
if (mdoc->meta.title) |
free(mdoc->meta.title); |
free(mdoc->meta.title); |
if (mdoc->meta.os) |
if (mdoc->meta.os) |
Line 276 mdoc_endparse(struct mdoc *m) |
|
Line 279 mdoc_endparse(struct mdoc *m) |
|
|
|
/* |
/* |
* Main parse routine. Parses a single line -- really just hands off to |
* Main parse routine. Parses a single line -- really just hands off to |
* the macro (parsemacro()) or text parser (parsetext()). |
* the macro (mdoc_pmacro()) or text parser (mdoc_ptext()). |
*/ |
*/ |
int |
int |
mdoc_parseln(struct mdoc *m, int ln, char *buf) |
mdoc_parseln(struct mdoc *m, int ln, char *buf) |
Line 285 mdoc_parseln(struct mdoc *m, int ln, char *buf) |
|
Line 288 mdoc_parseln(struct mdoc *m, int ln, char *buf) |
|
if (MDOC_HALT & m->flags) |
if (MDOC_HALT & m->flags) |
return(0); |
return(0); |
|
|
return('.' == *buf ? parsemacro(m, ln, buf) : |
m->flags |= MDOC_NEWLINE; |
parsetext(m, ln, buf)); |
return('.' == *buf ? |
|
mdoc_pmacro(m, ln, buf) : |
|
mdoc_ptext(m, ln, buf)); |
} |
} |
|
|
|
|
|
|
mdoc_macro(struct mdoc *m, enum mdoct tok, |
mdoc_macro(struct mdoc *m, enum mdoct tok, |
int ln, int pp, int *pos, char *buf) |
int ln, int pp, int *pos, char *buf) |
{ |
{ |
|
|
assert(tok < MDOC_MAX); |
assert(tok < MDOC_MAX); |
/* |
|
* If we're in the prologue, deny "body" macros. Similarly, if |
/* If we're in the body, deny prologue calls. */ |
* we're in the body, deny prologue calls. |
|
*/ |
|
if (MDOC_PROLOGUE & mdoc_macros[tok].flags && |
if (MDOC_PROLOGUE & mdoc_macros[tok].flags && |
MDOC_PBODY & m->flags) |
MDOC_PBODY & m->flags) |
return(mdoc_perr(m, ln, pp, EPROLBODY)); |
return(mdoc_perr(m, ln, pp, EPROLBODY)); |
|
|
|
/* If we're in the prologue, deny "body" macros. */ |
|
|
if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && |
if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && |
! (MDOC_PBODY & m->flags)) |
! (MDOC_PBODY & m->flags)) { |
return(mdoc_perr(m, ln, pp, EBODYPROL)); |
if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL)) |
|
return(0); |
|
if (NULL == m->meta.title) |
|
m->meta.title = mandoc_strdup("unknown"); |
|
if (NULL == m->meta.vol) |
|
m->meta.vol = mandoc_strdup("local"); |
|
if (NULL == m->meta.os) |
|
m->meta.os = mandoc_strdup("local"); |
|
if (0 == m->meta.date) |
|
m->meta.date = time(NULL); |
|
m->flags |= MDOC_PBODY; |
|
} |
|
|
return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); |
return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); |
} |
} |
Line 437 node_alloc(struct mdoc *m, int line, int pos, |
|
Line 454 node_alloc(struct mdoc *m, int line, int pos, |
|
p->pos = pos; |
p->pos = pos; |
p->tok = tok; |
p->tok = tok; |
p->type = type; |
p->type = type; |
|
if (MDOC_NEWLINE & m->flags) |
|
p->flags |= MDOC_LINE; |
|
m->flags &= ~MDOC_NEWLINE; |
return(p); |
return(p); |
} |
} |
|
|
Line 518 mdoc_elem_alloc(struct mdoc *m, int line, int pos, |
|
Line 537 mdoc_elem_alloc(struct mdoc *m, int line, int pos, |
|
} |
} |
|
|
|
|
static int |
int |
pstring(struct mdoc *m, int line, int pos, const char *p, size_t len) |
mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) |
{ |
{ |
struct mdoc_node *n; |
struct mdoc_node *n; |
size_t sv; |
size_t sv, len; |
|
|
n = node_alloc(m, line, pos, -1, MDOC_TEXT); |
len = strlen(p); |
|
|
|
n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT); |
n->string = mandoc_malloc(len + 1); |
n->string = mandoc_malloc(len + 1); |
sv = strlcpy(n->string, p, len + 1); |
sv = strlcpy(n->string, p, len + 1); |
|
|
Line 533 pstring(struct mdoc *m, int line, int pos, const char |
|
Line 554 pstring(struct mdoc *m, int line, int pos, const char |
|
|
|
if ( ! node_append(m, n)) |
if ( ! node_append(m, n)) |
return(0); |
return(0); |
|
|
m->next = MDOC_NEXT_SIBLING; |
m->next = MDOC_NEXT_SIBLING; |
return(1); |
return(1); |
} |
} |
|
|
|
|
int |
|
mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p) |
|
{ |
|
|
|
return(pstring(m, line, pos, p, strlen(p))); |
|
} |
|
|
|
|
|
void |
void |
mdoc_node_free(struct mdoc_node *p) |
mdoc_node_free(struct mdoc_node *p) |
{ |
{ |
|
|
if (p->parent) |
|
p->parent->nchild--; |
|
if (p->string) |
if (p->string) |
free(p->string); |
free(p->string); |
if (p->args) |
if (p->args) |
Line 560 mdoc_node_free(struct mdoc_node *p) |
|
Line 572 mdoc_node_free(struct mdoc_node *p) |
|
} |
} |
|
|
|
|
void |
static void |
mdoc_node_freelist(struct mdoc_node *p) |
mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n) |
{ |
{ |
|
|
if (p->child) |
/* Adjust siblings. */ |
mdoc_node_freelist(p->child); |
|
if (p->next) |
|
mdoc_node_freelist(p->next); |
|
|
|
|
if (n->prev) |
|
n->prev->next = n->next; |
|
if (n->next) |
|
n->next->prev = n->prev; |
|
|
|
/* Adjust parent. */ |
|
|
|
if (n->parent) { |
|
n->parent->nchild--; |
|
if (n->parent->child == n) |
|
n->parent->child = n->prev ? n->prev : n->next; |
|
} |
|
|
|
/* Adjust parse point, if applicable. */ |
|
|
|
if (m && m->last == n) { |
|
if (n->prev) { |
|
m->last = n->prev; |
|
m->next = MDOC_NEXT_SIBLING; |
|
} else { |
|
m->last = n->parent; |
|
m->next = MDOC_NEXT_CHILD; |
|
} |
|
} |
|
|
|
if (m && m->first == n) |
|
m->first = NULL; |
|
} |
|
|
|
|
|
void |
|
mdoc_node_delete(struct mdoc *m, struct mdoc_node *p) |
|
{ |
|
|
|
while (p->child) { |
|
assert(p->nchild); |
|
mdoc_node_delete(m, p->child); |
|
} |
assert(0 == p->nchild); |
assert(0 == p->nchild); |
|
|
|
mdoc_node_unlink(m, p); |
mdoc_node_free(p); |
mdoc_node_free(p); |
} |
} |
|
|
Line 579 mdoc_node_freelist(struct mdoc_node *p) |
|
Line 628 mdoc_node_freelist(struct mdoc_node *p) |
|
* control character. |
* control character. |
*/ |
*/ |
static int |
static int |
parsetext(struct mdoc *m, int line, char *buf) |
mdoc_ptext(struct mdoc *m, int line, char *buf) |
{ |
{ |
int i, j; |
int i; |
char sv; |
|
|
|
|
/* Ignore bogus comments. */ |
|
|
|
if ('\\' == buf[0] && '.' == buf[1] && '\"' == buf[2]) |
|
return(mdoc_pwarn(m, line, 0, EBADCOMMENT)); |
|
|
|
/* No text before an initial macro. */ |
|
|
if (SEC_NONE == m->lastnamed) |
if (SEC_NONE == m->lastnamed) |
return(mdoc_perr(m, line, 0, ETEXTPROL)); |
return(mdoc_perr(m, line, 0, ETEXTPROL)); |
|
|
/* |
|
* If in literal mode, then pass the buffer directly to the |
|
* back-end, as it should be preserved as a single term. |
|
*/ |
|
|
|
|
/* Literal just gets pulled in as-is. */ |
|
|
if (MDOC_LITERAL & m->flags) |
if (MDOC_LITERAL & m->flags) |
return(mdoc_word_alloc(m, line, 0, buf)); |
return(mdoc_word_alloc(m, line, 0, buf)); |
|
|
/* Disallow blank/white-space lines in non-literal mode. */ |
/* Check for a blank line, which may also consist of spaces. */ |
|
|
for (i = 0; ' ' == buf[i]; i++) |
for (i = 0; ' ' == buf[i]; i++) |
/* Skip leading whitespace. */ ; |
/* Skip to first non-space. */ ; |
|
|
if ('\0' == buf[i]) |
if ('\0' == buf[i]) { |
return(mdoc_perr(m, line, 0, ENOBLANK)); |
if ( ! mdoc_pwarn(m, line, 0, ENOBLANK)) |
|
return(0); |
|
|
/* |
/* |
* Break apart a free-form line into tokens. Spaces are |
* Insert a `Pp' in the case of a blank line. Technically, |
* stripped out of the input. |
* blank lines aren't allowed, but enough manuals assume this |
*/ |
* behaviour that we want to work around it. |
|
*/ |
|
if ( ! mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL)) |
|
return(0); |
|
|
for (j = i; buf[i]; i++) { |
m->next = MDOC_NEXT_SIBLING; |
if (' ' != buf[i]) |
return(1); |
continue; |
} |
|
|
/* Escaped whitespace. */ |
/* |
if (i && ' ' == buf[i] && '\\' == buf[i - 1]) |
* Warn if the last un-escaped character is whitespace. Then |
continue; |
* strip away the remaining spaces (tabs stay!). |
|
*/ |
|
|
sv = buf[i]; |
i = (int)strlen(buf); |
buf[i++] = '\0'; |
assert(i); |
|
|
if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j))) |
if (' ' == buf[i - 1] || '\t' == buf[i - 1]) { |
return(0); |
if (i > 1 && '\\' != buf[i - 2]) |
|
|
/* Trailing whitespace? Check at overwritten byte. */ |
|
|
|
if (' ' == sv && '\0' == buf[i]) |
|
if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS)) |
if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS)) |
return(0); |
return(0); |
|
|
for ( ; ' ' == buf[i]; i++) |
for (--i; i && ' ' == buf[i]; i--) |
/* Skip trailing whitespace. */ ; |
/* Spin back to non-space. */ ; |
|
|
j = i; |
/* Jump ahead of escaped whitespace. */ |
|
i += '\\' == buf[i] ? 2 : 1; |
|
|
/* Trailing whitespace? */ |
buf[i] = '\0'; |
|
|
if (' ' == buf[i - 1] && '\0' == buf[i]) |
|
if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS)) |
|
return(0); |
|
|
|
if ('\0' == buf[i]) |
|
break; |
|
} |
} |
|
|
if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j))) |
/* Allocate the whole word. */ |
|
|
|
if ( ! mdoc_word_alloc(m, line, 0, buf)) |
return(0); |
return(0); |
|
|
m->next = MDOC_NEXT_SIBLING; |
/* |
|
* End-of-sentence check. If the last character is an unescaped |
|
* EOS character, then flag the node as being the end of a |
|
* sentence. The front-end will know how to interpret this. |
|
*/ |
|
|
|
assert(i); |
|
|
|
if (mandoc_eos(buf, (size_t)i)) |
|
m->last->flags |= MDOC_EOS; |
|
|
return(1); |
return(1); |
} |
} |
|
|
|
|
|
|
static int |
static int |
macrowarn(struct mdoc *m, int ln, const char *buf) |
macrowarn(struct mdoc *m, int ln, const char *buf) |
{ |
{ |
if ( ! (MDOC_IGN_MACRO & m->pflags)) |
if ( ! (MDOC_IGN_MACRO & m->pflags)) |
return(mdoc_verr(m, ln, 0, |
return(mdoc_verr(m, ln, 0, "unknown macro: %s%s", |
"unknown macro: %s%s", |
|
buf, strlen(buf) > 3 ? "..." : "")); |
buf, strlen(buf) > 3 ? "..." : "")); |
return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s", |
return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s", |
buf, strlen(buf) > 3 ? "..." : "")); |
buf, strlen(buf) > 3 ? "..." : "")); |
Line 669 macrowarn(struct mdoc *m, int ln, const char *buf) |
|
Line 726 macrowarn(struct mdoc *m, int ln, const char *buf) |
|
* character. |
* character. |
*/ |
*/ |
int |
int |
parsemacro(struct mdoc *m, int ln, char *buf) |
mdoc_pmacro(struct mdoc *m, int ln, char *buf) |
{ |
{ |
int i, j, c; |
enum mdoct tok; |
char mac[5]; |
int i, j, sv; |
|
char mac[5]; |
|
|
/* Empty lines are ignored. */ |
/* Empty lines are ignored. */ |
|
|
Line 691 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 749 parsemacro(struct mdoc *m, int ln, char *buf) |
|
return(1); |
return(1); |
} |
} |
|
|
|
sv = i; |
|
|
/* Copy the first word into a nil-terminated buffer. */ |
/* Copy the first word into a nil-terminated buffer. */ |
|
|
for (j = 0; j < 4; j++, i++) { |
for (j = 0; j < 4; j++, i++) { |
Line 714 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 774 parsemacro(struct mdoc *m, int ln, char *buf) |
|
return(1); |
return(1); |
} |
} |
|
|
if (MDOC_MAX == (c = mdoc_hash_find(mac))) { |
if (MDOC_MAX == (tok = mdoc_hash_find(mac))) { |
if ( ! macrowarn(m, ln, mac)) |
if ( ! macrowarn(m, ln, mac)) |
goto err; |
goto err; |
return(1); |
return(1); |
Line 725 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 785 parsemacro(struct mdoc *m, int ln, char *buf) |
|
while (buf[i] && ' ' == buf[i]) |
while (buf[i] && ' ' == buf[i]) |
i++; |
i++; |
|
|
/* Trailing whitespace? */ |
/* |
|
* Trailing whitespace. Note that tabs are allowed to be passed |
|
* into the parser as "text", so we only warn about spaces here. |
|
*/ |
|
|
if ('\0' == buf[i] && ' ' == buf[i - 1]) |
if ('\0' == buf[i] && ' ' == buf[i - 1]) |
if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS)) |
if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS)) |
Line 735 parsemacro(struct mdoc *m, int ln, char *buf) |
|
Line 798 parsemacro(struct mdoc *m, int ln, char *buf) |
|
* Begin recursive parse sequence. Since we're at the start of |
* Begin recursive parse sequence. Since we're at the start of |
* the line, we don't need to do callable/parseable checks. |
* the line, we don't need to do callable/parseable checks. |
*/ |
*/ |
if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) |
if ( ! mdoc_macro(m, tok, ln, sv, &i, buf)) |
goto err; |
goto err; |
|
|
return(1); |
return(1); |